基于信道估计的LDPC仿真算法研究  

Simulation of LDPC Decoding Algorithm Based on Channel Estimation

作　　者：方美彦[1] 刘粉山[1] 王继康[2] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027 [2]中国科学技术大学自动化系,安徽合肥230027

出　　处：《计算机仿真》2007年第3期90-93,122,共5页Computer Simulation

摘　　要：研究有记忆信道上的LDPC译码算法,对高速数字通信系统具有重要意义。目前运用于有记忆信道上的LDPC迭代译码算法,如基于信道估计的BP迭代译码算法等,都存在算法复杂度较高、运算量较大的问题。针对隐马尔可夫噪声信道,首次将最小和(min-sum)算法引入到基于噪声软判决和信道估计的LDPC迭代译码算法,利用函数特性有效降低算法复杂度、减少运算量。仿真结果表明,此算法的性能不仅优于不考虑信道记忆特性的一般LDPC的迭代译码算法,也优于基于噪声硬判决和信道估计的BP迭代译码算法,在性能损失较小情况下,于译码性能和算法复杂度之间找到了一个很好的折衷,对实时通信系统具有重要意义。LDPC decoding algorithm on channels with memory is of great significance to high speed digital communications. Currently used low -density parity -check codes iterative decoding algorithms on channels with memory, such as Belief - Propagation iterative decoding algorithm based on channel estimation, are commonly complex and thus requiring large amount of calculation. In this paper, a rain - sum algorithm is firstly introduced into LDPC iterarive decoding algorithm based on soft noise decisions and channel estimation on hidden Markov noisy channels. Algorithm and calculation complexity are effectively reduced by using the functions＇characteristics. Simulation results show that the algorithm not only excels conventional Belief - Propagation iterative decoding algorithms that do not concern channel memory, but also outperforms iterative decoding algorithms based on hard noise decisions and channel estimation. So a good compromise is made between decoding performance and algorithm complexity with low performance loss, which is useful real time communications.

关 键 词：隐马尔可夫模型 置信传播 最小和算法 

分 类 号：TN911.22[电子电信—通信与信息系统]